mikhail/llama.cpp-mtp-turboquant
1
0
Fork 0
Code Issues Pull Requests Actions 30 Packages Projects Releases Wiki Activity

model: support GLM-OCR (#19677)

Browse Source 
* model: support GLM-OCR

* Update convert_hf_to_gguf.py

Co-authored-by: Sigbjørn Skjæret <sigbjorn.skjaeret@scala.com>

---------

Co-authored-by: Sigbjørn Skjæret <sigbjorn.skjaeret@scala.com>
This commit is contained in:
Xuan-Son Nguyen
2026-02-18 17:51:40 +01:00
committed by GitHub
parent e99f1083a0
commit eeef3cfced
8 changed files with 122 additions and 43 deletions
Download Patch File Download Diff File Expand all files Collapse all files
gguf-py/gguf/constants.py
+7
View File
@@ -2660,6 +2660,13 @@ MODEL_TENSORS: dict[MODEL_ARCH, list[MODEL_TENSOR]] = {
MODEL_TENSOR.FFN_UP,
MODEL_TENSOR.ATTN_POST_NORM,
MODEL_TENSOR.FFN_POST_NORM,
# NextN/MTP tensors - preserved but unused
MODEL_TENSOR.NEXTN_EH_PROJ,
MODEL_TENSOR.NEXTN_EMBED_TOKENS,
MODEL_TENSOR.NEXTN_ENORM,
MODEL_TENSOR.NEXTN_HNORM,
MODEL_TENSOR.NEXTN_SHARED_HEAD_HEAD,
MODEL_TENSOR.NEXTN_SHARED_HEAD_NORM,
],
MODEL_ARCH.GLM4_MOE: [
MODEL_TENSOR.TOKEN_EMBD,
gguf-py/gguf/tensor_mapping.py
+2
View File
@@ -1404,6 +1404,7 @@ class TensorNameMap:
MODEL_TENSOR.V_ENC_ATTN_Q_NORM: (
"vision_tower.vision_model.encoder.layers.{bid}.attn.q_norm", # InternVL
"model.vision_tower.encoder.layer.{bid}.attention.q_norm", # Intern-S1
"visual.blocks.{bid}.attn.q_norm", # GLM-OCR
),
MODEL_TENSOR.V_ENC_ATTN_K: (
@@ -1422,6 +1423,7 @@ class TensorNameMap:
MODEL_TENSOR.V_ENC_ATTN_K_NORM: (
"vision_tower.vision_model.encoder.layers.{bid}.attn.k_norm", # InternVL
"model.vision_tower.encoder.layer.{bid}.attention.k_norm", # Intern-S1
"visual.blocks.{bid}.attn.k_norm", # GLM-OCR
),
MODEL_TENSOR.V_ENC_ATTN_V: (